Akash Agarwal

Engineered a robust algorithm for identifying VoIP calls from raw internet data using Deep Packet Inspection, strengthening the product portfolio with advanced data analysis capabilities. This innovation delivered an outstanding precision rate of 90%. Designed RESTful API endpoints using django to filter internet subscribers based on user-selectable activity filters for backend microservice. Architectured and implemented a centralized product configuration and management system using ConfD and data modeling languages like YANG, which simplified configuration process from 182 steps to 40 steps resulting in fastened customer experience. Constructed a generic ETL (Extract, Transform, Load) pipeline with multiprocessing capabilities to efficiently detect social media uploads from raw internet data using data manipulation, ensuring seamless data flow from inception to algorithm-based upload detection. This enhanced the product portfolio by 20%. Architectured and developed a unified cleanup module for efficiently pruning of records from the database using user-defined date ranges. This contributed to a significant reduction of up to 25% in database storage requirement. Designed and constructed a data processing pipeline framework with Luigi for batch job, workflow management and dependency resolution. Led feature development and acted as Scrum Master for an Agile-based Scrum team, overseeing the synchronization of SQL tables across multiple database clusters. Successfully mentored individuals, facilitating their comprehensive understanding of the codebase. Employed effective guidance and communication strategies to ensure a seamless integration process, contributing to the team’s overall proficiency and success. Resolved critical issues and queries leading to customer retention. Served as the primary customer contact point and provided direct debugging and enhancement support for the backend team, ensuring a seamless customer experience.

Conducted a feasibility study on Voice call detection and implemented an algorithm for correlating FaceTime calls within subscriber data. Structured unit tests for VoIP Call detection using Python’s unittest library. Conducted VoIP call packet capture analysis using Wireshark to derive key metrics for correlating calls.